You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I searched existing issues before opening this one
Expected behavior
Two containers connected to the same internal docker network should be able to communicate to each other.
Actual behavior
Pinging one container from inside another works fine but establishing a TCP connection (in this case connecting to an nginx server) fails with a "Host is unreachable" error (or in case of RHEL 8 on AWS a "No route to host" error).
Steps to reproduce the behavior
Create an AWS EC2 instance "Red Hat Enterprise Linux 8 (HVM), SSD Volume Type - ami-06ec8443c2a35b0ba (64-bit x86)"
I originally encountered this issue on CentOS 8 running on a physical server.
I also noticed that the bridge network device (in this case br-c81932d5907d) is not assigned to the docker zone in firewalld (or any zone for that matter). It uses whichever zone is set as default.
Expected behavior
Two containers connected to the same
internal
docker network should be able to communicate to each other.Actual behavior
Pinging one container from inside another works fine but establishing a TCP connection (in this case connecting to an nginx server) fails with a "Host is unreachable" error (or in case of RHEL 8 on AWS a "No route to host" error).
Steps to reproduce the behavior
Create an AWS EC2 instance "Red Hat Enterprise Linux 8 (HVM), SSD Volume Type - ami-06ec8443c2a35b0ba (64-bit x86)"
Install
firewalld
Install
docker
Create
internal
network and two containersAttach to
client
and try to connect tonginx
As you can see pinging
nginx
works fineOpening a TCP connection to port 80 however doesn't work
Output of
docker version
:Output of
docker info
:Additional environment details (AWS, VirtualBox, physical, etc.)
I originally encountered this issue on CentOS 8 running on a physical server.
I also noticed that the bridge network device (in this case
br-c81932d5907d
) is not assigned to thedocker
zone in firewalld (or any zone for that matter). It uses whichever zone is set as default.nftables
trace (nft add chain bridge filter trace_chain { type filter hook forward priority -301\; }
) while runningnc -v 172.19.0.2
:The text was updated successfully, but these errors were encountered: